· 个股论点
出口管制是最大的风险。但他们日本供应链中的两大竞争对手刚刚被禁止出口。因此，现在 $AXTI 以及通过一步跳板的中国，控制了磷化铟(InP)衬底和材料的流向。我认为许多分析师在建模时忽略的一点是，这并非关于扩大产量或数量，而是基于供应有限的关键资源进行定价。这就是博弈论在资源分配中的应用，类似于高带宽内存(HBM)的情况，但不同的是主要只有一个玩家。磷化铟(InP)的潜在市场规模(TAM)此前仅为数亿美元，因为它是一种小众的电信产品。现在，$GOOGL 愿意支付数十亿美元，仅为了确保其 Ironwood 项目不会停滞。其他超大规模云服务商也是如此。当各方开始争夺配额时，TAM 将从数亿美元跃升至数十亿甚至 100 亿美元以上。而 $AXTI 在两个不同瓶颈环节占据垄断地位，市值却仅为 12 亿美元左右。例如，价格增加 3000% 在专用集成电路(ASIC)集群的物料清单(BOM)价值中仅占百分之几。在出口管制下，你不需要生产多 3000% 的材料，只需提高价格即可。
Export controls is the biggest risk. But their two biggest competitors over in the Japanese supply chain just got export control banned. So now $AXTI and by one hop, China, controls the flow of InP substrates and materials. I think what a lot of analysts miss when modelling this is that it's not about scaling up production or quantity, it's about pricing based on a critical resource with finite supply. This is where game theory comes into play about resource allocation like what's going on with HBM, but instead of multiple players you mainly have one. The TAM of InP earlier few hundred million because it was a niche telecom product. Now, $GOOGL would happily pay billions just so their ironwood program doesn't stall out. Same with every other hyperscaler. Then TAM goes from few hundred million to few billions or $10B+ when they start fighting for allocations. And MC is still only $1.2Bish for a monopolistic position in two different bottlenecks. A 3000%+ increase for example in price would only be a few percent BOM value in an asic cluster. You don't need to produce 3000% more material with export controls, just increase the prices.